The Experts below are selected from a list of 37128 Experts worldwide ranked by ideXlab platform
Wusheng Lu - One of the best experts on this subject based on the ideXlab platform.
-
roundoff noise minimization for 2 d separable denominator digital filters using jointly optimal high order error feedback and realization
International Symposium on Circuits and Systems, 2017Co-Authors: Takao Hinamoto, Wusheng LuAbstract:The joint optimization problem of high-order error feedback and realization for minimizing roundoff noise at the filter output subject to-scaling constraints for two-dimensional (2-D) separable-denominator digital filters is investigated. Linear algebraic techniques that convert the problem at hand into an unconstrained optimization problem are explored, and an efficient quasi-Newton Algorithm is then applied to solve the unconstrained optimization problem iteratively. Closed-form formulas for fast and accurate gradient evaluation are derived. A numerical example is presented to demonstrate the validity and effectiveness of the proposed technique.
-
jointly optimal high order error feedback and realization for roundoff noise minimization in 2 d state space digital filters
European Conference on Circuit Theory and Design, 2013Co-Authors: Takao Hinamoto, Wusheng LuAbstract:The joint optimization problem of high-order error feedback and realization for minimizing roundoff noise at filter output subject to l2-scaling constraints is investigated for two-dimensional (2-D) state-space digital filters. Linear algebraic techniques that convert the problem at hand into an unconstrained optimization problem are explored, and an efficient quasi-Newton Algorithm is then applied to solve the unconstrained optimization problem iteratively. In this connection, closed-form formulas are derived for fast and accurate gradient evaluation. Finally a numerical example is presented to illustrate the validity and effectiveness of the proposed Algorithm.
-
jointly optimized high order error feedback and realization for roundoff noise minimizaion in state space digital filters
International Midwest Symposium on Circuits and Systems, 2011Co-Authors: Takao Hinamoto, Wusheng LuAbstract:The joint optimization problem of high-order error feedback and realization for state-space digital filters to minimize the effects of roundoff noise at the filter output subject to l 2 -scaling constraints is investigated. The problem is converted into an unconstrained optimization problem by using linear-algebraic techniques. The unconstrained optimization problem at hand is then solved iteratively by applying an efficient quasi-Newton Algorithm with closed-form formulas for key gradient evaluation. Finally, a numerical example is presented to illustrate the utility of the proposed technique.
-
separate joint optimization of error feedback and realization for roundoff noise minimization in a class of 2 d state space digital filters
Multidimensional Systems and Signal Processing, 2007Co-Authors: Takao Hinamoto, Toiru Oumi, Wusheng LuAbstract:Techniques for the separate/joint optimization of error-feedback and realization are developed to minimize the roundoff noise subject to l 2-norm dynamic-range scaling constraints for a class of 2-D state-space digital filters. In the joint optimization, the problem at hand is converted into an unconstrained optimization problem by using linear-algebraic techniques. The unconstrained problem obtained is then solved by applying an efficient quasi-Newton Algorithm. A numerical example is presented to illustrate the utility of the proposed techniques.
-
jointly optimized error feedback and realization for roundoff noise minimization in a class of 2 d state space digital filters
International Symposium on Signals Circuits and Systems, 2007Co-Authors: Takao Hinamoto, Toiru Oumi, Wusheng LuAbstract:The separate/joint optimization techniques of error-feedback and realization are explored so as to reduce or minimize the effects of roundoff noise subject to l2-scaling constraints for a class of 2D state-space digital filters. First, the joint optimization problem at hand is converted into an unconstrained optimization problem by using linear-algebraic techniques. Next, the resultant unconstrained problem is solved iteratively by applying an efficient quasi-Newton Algorithm. Finally, a numerical example is presented to illustrate the utility of the proposed techniques.
Takao Hinamoto - One of the best experts on this subject based on the ideXlab platform.
-
roundoff noise minimization for 2 d separable denominator digital filters using jointly optimal high order error feedback and realization
International Symposium on Circuits and Systems, 2017Co-Authors: Takao Hinamoto, Wusheng LuAbstract:The joint optimization problem of high-order error feedback and realization for minimizing roundoff noise at the filter output subject to-scaling constraints for two-dimensional (2-D) separable-denominator digital filters is investigated. Linear algebraic techniques that convert the problem at hand into an unconstrained optimization problem are explored, and an efficient quasi-Newton Algorithm is then applied to solve the unconstrained optimization problem iteratively. Closed-form formulas for fast and accurate gradient evaluation are derived. A numerical example is presented to demonstrate the validity and effectiveness of the proposed technique.
-
jointly optimal high order error feedback and realization for roundoff noise minimization in 2 d state space digital filters
European Conference on Circuit Theory and Design, 2013Co-Authors: Takao Hinamoto, Wusheng LuAbstract:The joint optimization problem of high-order error feedback and realization for minimizing roundoff noise at filter output subject to l2-scaling constraints is investigated for two-dimensional (2-D) state-space digital filters. Linear algebraic techniques that convert the problem at hand into an unconstrained optimization problem are explored, and an efficient quasi-Newton Algorithm is then applied to solve the unconstrained optimization problem iteratively. In this connection, closed-form formulas are derived for fast and accurate gradient evaluation. Finally a numerical example is presented to illustrate the validity and effectiveness of the proposed Algorithm.
-
jointly optimized high order error feedback and realization for roundoff noise minimizaion in state space digital filters
International Midwest Symposium on Circuits and Systems, 2011Co-Authors: Takao Hinamoto, Wusheng LuAbstract:The joint optimization problem of high-order error feedback and realization for state-space digital filters to minimize the effects of roundoff noise at the filter output subject to l 2 -scaling constraints is investigated. The problem is converted into an unconstrained optimization problem by using linear-algebraic techniques. The unconstrained optimization problem at hand is then solved iteratively by applying an efficient quasi-Newton Algorithm with closed-form formulas for key gradient evaluation. Finally, a numerical example is presented to illustrate the utility of the proposed technique.
-
separate joint optimization of error feedback and realization for roundoff noise minimization in a class of 2 d state space digital filters
Multidimensional Systems and Signal Processing, 2007Co-Authors: Takao Hinamoto, Toiru Oumi, Wusheng LuAbstract:Techniques for the separate/joint optimization of error-feedback and realization are developed to minimize the roundoff noise subject to l 2-norm dynamic-range scaling constraints for a class of 2-D state-space digital filters. In the joint optimization, the problem at hand is converted into an unconstrained optimization problem by using linear-algebraic techniques. The unconstrained problem obtained is then solved by applying an efficient quasi-Newton Algorithm. A numerical example is presented to illustrate the utility of the proposed techniques.
-
jointly optimized error feedback and realization for roundoff noise minimization in a class of 2 d state space digital filters
International Symposium on Signals Circuits and Systems, 2007Co-Authors: Takao Hinamoto, Toiru Oumi, Wusheng LuAbstract:The separate/joint optimization techniques of error-feedback and realization are explored so as to reduce or minimize the effects of roundoff noise subject to l2-scaling constraints for a class of 2D state-space digital filters. First, the joint optimization problem at hand is converted into an unconstrained optimization problem by using linear-algebraic techniques. Next, the resultant unconstrained problem is solved iteratively by applying an efficient quasi-Newton Algorithm. Finally, a numerical example is presented to illustrate the utility of the proposed techniques.
Zaid Harchaoui - One of the best experts on this subject based on the ideXlab platform.
-
a generic quasi Newton Algorithm for faster gradient based optimization
arXiv: Machine Learning, 2017Co-Authors: Hongzhou Lin, Julien Mairal, Zaid HarchaouiAbstract:We propose a generic approach to accelerate gradient-based optimization Algorithms with quasi-Newton principles. The proposed scheme, called QuickeNing, can be applied to incremental first-order methods such as stochastic variance-reduced gradient (SVRG) or incremental surrogate optimization (MISO). It is also compatible with composite objectives, meaning that it has the ability to provide exactly sparse solutions when the objective involves a sparsity-inducing regularization. QuickeNing relies on limited-memory BFGS rules, making it appropriate for solving high-dimensional optimization problems. Besides, it enjoys a worst-case linear convergence rate for strongly convex problems. We present experimental results where QuickeNing gives significant improvements over competing methods for solving large-scale high-dimensional machine learning problems.
-
quickening a generic quasi Newton Algorithm for faster gradient based optimization
2016Co-Authors: Hongzhou Lin, Julien Mairal, Zaid HarchaouiAbstract:We propose an approach to accelerate gradient-based optimization Algorithms by giving them the ability to exploit curvature information using quasi-Newton update rules. The proposed scheme, called QuickeNing, is generic and can be applied to a large class of first-order methods such as incremental and block-coordinate Algorithms; it is also compatible with composite objectives, meaning that it has the ability to provide exactly sparse solutions when the objective involves a sparsity-inducing regularization. QuickeNing relies on limited-memory BFGS rules, making it appropriate for solving high-dimensional optimization problems; with no line-search, it is also simple to use and to implement. Besides, it enjoys a worst-case linear convergence rate for strongly convex problems. We present experimental results where QuickeNing gives significant improvements over competing methods for solving large-scale high-dimensional machine learning problems.
Qingchuan Zhang - One of the best experts on this subject based on the ideXlab platform.
-
interpolation bias for the inverse compositional gauss Newton Algorithm in digital image correlation
Optics and Lasers in Engineering, 2018Co-Authors: Yong Su, Xiaohai Xu, Qingchuan Zhang, Shangquan WuAbstract:Abstract It is believed that the classic forward additive Newton–Raphson (FA-NR) Algorithm and the recently introduced inverse compositional Gauss–Newton (IC-GN) Algorithm give rise to roughly equal interpolation bias. Questioning the correctness of this statement, this paper presents a thorough analysis of interpolation bias for the IC-GN Algorithm. A theoretical model is built to analytically characterize the dependence of interpolation bias upon speckle image, target image interpolation, and reference image gradient estimation. The interpolation biases of the FA-NR Algorithm and the IC-GN Algorithm can be significantly different, whose relative difference can exceed 80%. For the IC-GN Algorithm, the gradient estimator can strongly affect the interpolation bias; the relative difference can reach 178%. Since the mean bias errors are insensitive to image noise, the theoretical model proposed remains valid in the presence of noise. To provide more implementation details, source codes are uploaded as a supplement.
-
high efficiency and high accuracy digital image correlation for three dimensional measurement
Optics and Lasers in Engineering, 2015Co-Authors: Teng Cheng, Yong Su, Xiaohai Xu, Yong Zhang, Qingchuan ZhangAbstract:Abstract The computational efficiency and measurement accuracy of the digital image correlation (DIC) have become more and more important in recent years. For the three-dimensional DIC (3D-DIC), these issues are much more serious. First, there are two cameras employed which increases the computational amount several times. Second, because of the differences in view angles, the must-do stereo correspondence between the left and right images is equivalently a non-uniform deformation, and cannot be weakened by increasing the sampling frequency of digital cameras. This work mainly focuses on the efficiency and accuracy of 3D-DIC. The inverse compositional Gauss–Newton Algorithm (IC-GN 2 ) with the second-order shape function is firstly proposed. Because it contains the second-order displacement gradient terms, the measurement accuracy for the non-uniform deformation thus can be improved significantly, which is typically one order higher than the first-order shape function combined with the IC-GN Algorithm (IC-GN 1 ), and 2 times faster than the second-order shape function combined with the forward additive Gauss–Newton Algorithm (FA-GN 2 ). Then, based on the features of the IC-GN 1 and IC-GN 2 Algorithms, a high-efficiency and high-accuracy measurement strategy for 3D-DIC is proposed in the end.
Shinichi Koike - One of the best experts on this subject based on the ideXlab platform.
-
adaptive step size biphase error Newton Algorithm
International Symposium on Intelligent Signal Processing and Communication Systems, 2018Co-Authors: Shinichi KoikeAbstract:This paper first derives Biphase Error Algorithm (BiPhEA) using a “biphase function” and then proposes Adaptive Step-Size Biphase Error Newton Algorithm (ABENA) for adaptive filters in the complex-number domain with a Gaussian regressor. We present a stochastic model called Contaminated Gaussian Noise (CGN) for impulsive observation noise found at the filter output. To improve the filter convergence for a correlated regressor, we combine the BiPhEA with an estimate of the inverse covariance matrix of the regressor calculated using the Newton's method. We further propose a new stable Adaptive Step-Size (ASS) control Algorithm. Performance analysis of the ABENA is developed for theoretically calculating transient and steady-state convergence behavior. Through experiments with typical examples, we demonstrate faster convergence and high robustness of the proposed ABENA against the CGN. Good agreement between simulated and theoretical convergence curves validates the analysis.
-
analysis of adaptive step size correlation phase Newton Algorithm
International Symposium on Communications and Information Technologies, 2013Co-Authors: Shinichi KoikeAbstract:This paper first reviews Correlation Phase Algorithm (CPhiA) for adaptive filters in the complex-number domain with colored Gaussian reference inputs. Stochastic models are presented for two types of impulse noise intruding adaptive filters: one in observation noise and another at filter input. To improve the filter convergence speed for a strongly correlated filter reference input, we combine the CPhiA with simple recurrent calculation of the inverse covariance matrix of the filter input using the Newton's method (CPhi-Newton Algorithm). We further propose a new adaptive step-size (ASS) control Algorithm to be combined with the CPhi-Newton Algorithm. We develop performance analysis of the ASS-CPhi-Newton Algorithm to derive difference equations for theoretically calculating transient and steady-state convergence behavior. Through experiment with some examples, it is demonstrated that faster convergence and high robustness are realized with the proposed ASS-CPhi-Newton Algorithm in impulsive noise environments. Good agreement between simulated and theoretical convergence validates the analysis.
-
performance analysis of adaptive step size least mean modulus Newton Algorithm
International Symposium on Intelligent Signal Processing and Communication Systems, 2011Co-Authors: Shinichi KoikeAbstract:This paper first reviews least mean modulus-Newton (LMM-Newton) Algorithm for complex-domain adaptive filters. The LMM-Newton Algorithm is effective in making the convergence of an adaptive filter with a highly correlated input as fast as that for the LMM Algorithm with a White & Gaussian filter input. However, the filter convergence for the LMM-Newton Algorithm is still much slower than for the LMS Algorithm. Then, the paper introduces a generalized error modulus (“p-modulus”) and proposes a new adaptive step-size (ASS) control Algorithm to be combined with the LMM-Newton Algorithm to further improve the convergence speed. Analysis of the ASS-LMM-Newton Algorithm is developed for calculating transient and steady-state behavior. Through experiment with simulations and theoretical calculations of filter convergence, we find that the filter convergence is almost the same for any value of p of “p-modulus.” We demonstrate effectiveness of the proposed ASS-LMM-Newton Algorithm, while preserving the robustness of the LMM Algorithm against impulsive observation noise. Good agreement between simulated and theoretical convergence validates the analysis.